Sniffing and Epilepsy...Why?

Before I start the post

I put this up last week but my PI, (my boss for you non-scientists), sent me an urgent text to take it down ASAP. There's this huge issue in science called 'scooping.' This is where a scientist will see somebody else's work on a project that is similar to their own and race to publish it first, taking all of the credit. This happens far more often than you might think, especially among a group people who are all generally trying to help the world in some way. It's pretty weird and awful. My PI read the first paragraph and to her it looked like what I was saying was putting my work, as well as the other people in our lab, at risk of being scooped. She gave it a full read-through this weekend and gave it the OK. Most of the science and the methodologies that I write about here have actually been known and used for a long time. I wanted to write this thing because I'm really stoked about what we're working on and wanted to spread the word beyond the small circle of scientists that we interact with at talks and conferences. It turns out that you should at least ask somebody whether its alright before you launch their ideas out to wherever they go on the internet. Lesson learned. 
Now let's talk about this sniffing stuff.

I spent the last few months applying for a grant through the NIH to study...sniffing. 

If all goes as planned, some of your hard-earned tax dollars are going to directly to me to spend a couple of years investigating sniffing. Some scientists study the science of olfaction, or how your brain forms perceptions of odors from the chemicals that enter your nose. My research isn't so much about that. I'm literally going to study how people sniff stuff. 

If that isn't weird enough for you, I plan to work with epilepsy patients at Northwestern Hospital and ask them to smell a bunch of heinous odors while I record electrical signals from their brains and breathing waveforms from their noses.

So this obviously seems pretty insane. 

Neuroscientists should probably work on important problems like Alzheimer's Disease and MS. What reason could ever justify this trash?

My mentor, lab mates, and I believe that the results of this project could actually have pretty significant implications for understanding and treating epilepsy. In the rest of this post I'm going to walk you through our logic and hopefully prove to you that I haven't completely lost my mind. I'm going to start by explaining the basics of epilepsy.


Epilepsy and SUDEP

Epilepsy is a neurological disorder that is characterized by the recurrent tendency for one to have seizures. Seizures can come in many forms. For instance, someone having an absence seizure will appear to space out for a few seconds. These can be so subtle that many people who have absence seizures don't even realize that they have epilepsy for a long time. On the other hand, someone having a grand mal seizure will experience violent convulsions for up to about two minutes.

If you look someone's brain activity while they have a seizure, you will see that neurons in one area of the brain will start to fire more rapidly and synchronously than they ever normally would. In more severe seizures, this rampant abnormal activity can spread across the brain, sometimes reaching areas far from where it started. For example, in a grand mal seizure, the seizure activity reaches the motor cortex, which is responsible for many aspects of controlling movement. These neurons are mostly silent when you’re resting. But when you chose to move part of your body, a select group of these neurons will activate in a complex, finely-tuned pattern, sending signals down your spinal cord, precisely controlling which muscles to flex. However, when seizure activity sweeps across the motor cortex, tons of neurons are simultaneously activated, causing the erratic, unintentional muscle movements seen in grand mal seizures. 

The main takeaway here is that seizures disrupt the neural activity in the parts of the brain that they spread to and can initiate (or modify) the processes that neurons in those areas are involved in. 

Lots of great people are working on understanding seizures and epilepsy. I'm particularly excited about brain implants that can predict and disrupt seizures. Few people know that these exist and are actually inside some peoples' heads right now! (seriously awesome right?)  My work is related to SUDEP, a fatal complication of epilepsy that afflicts people with severe, medication-resistant cases. It is the leading cause of death for people with severe epilepsy.

SUDEP stands for Sudden Unexpected Death in EPilepsy. If you can guess by the name, we really have no idea how to predict SUDEP and we have just a vague idea of its cause. This is clearly a huge problem and we want to make a big dent in it. 

My brilliant supervisor had the idea that studying the neurons that control sniffing might actually help us understand SUDEP, and hopefully prevent it in the future. 

So why in tarnation would anyone think that?


Epilepsy and Breathing

One of the first key clues about the cause of SUDEP is evidence suggesting that these people may die because they stop breathing, which is soon followed by cardiac arrest [1,2]. This is a huge step forward but it is also rather unexpected.

More often than not, seizures begin in the temporal lobes of the brain, while breathing is controlled by a few small patches of neurons far away in the brainstem. Therefore, the first big question that we need to answer is, how would disrupted activity in these epileptic brain regions trigger neurons in a totally different area of the brain to make somebody stop breathing?

how tho?

Basically, for this to happen, there has to be an anatomical pathway linking epileptogenic neurons in the temporal lobe to respiratory neurons in the brainstem. I wasn't taught any such pathways in my neuroscience education, but that doesn't mean they don't exist. After some research, we came across some previous work that identified such a pathway in rats [3] and non-human primates [4], suggesting that we may have it in our brains as well. The amygdala, a brain region within each temporal lobe, contains a subregion called the central nucleus, which appears to have neurons that directly project to respiratory areas in the brainstem through the amygdalofugal pathway! Therefore, it seems plausible that when a seizure propagates to neurons in the central amygdala, this activity may reach respiratory neurons in the brainstem where it could disrupt — and maybe halt — ongoing breathing rhythms.

Now, anatomical evidence is good but it’s critical that we to test whether this connection matters. In other words, we need to experimentally test the hypothesis that activating neurons in the central amygdala causes people to stop breathing.

Testing this hypothesis is tricky because you can't just go around making people undergo brain surgery to stick wires in their amygdala and mess with their neurons. In the face of this completely reasonable constraint, many neuroscientists choose to do experiments on non-human animals. This is a totally valid approach but there’s always going to be uncertainty about whether something found in the brain of another species is going to be present in our brains as well. In the end we’re really going to have to test any hypothesis about humans directly in humans to be sure. Fortunately, there is actually a rare and special condition where we can do this. 


Doing Neuroscience with Epilepsy Patients

This kind of surgery has been done
since the 40's!

If you have debilitating, treatment-resistant epilepsy, sometimes neurologists will recommend that you undergo surgery to remove the part of the brain where your seizures originate. This is a difficult and involved procedure that can take days and sometimes weeks. During this treatment, electrodes are surgically implanted in the patient’s brain. Then, the patient is taken off their anti-seizure drugs and monitored 24/7, until they experience several clinical seizures. During the seizures, the electrodes in their brain record data that neurologists and technicians use to try to figure out where the seizures originate. Using up to around one hundred contact sites electrodes they create a 3D map of the epileptogenic site, or the seizure’s starting point, and follow where it spreads through the brain. Once the neurologists are confident that they have found the epileptogenic site, they perform another step before surgically removing it.

The electrodes implanted in the patient’s brain can stimulate as well as record. With electrical neurologists can perturb areas around the resection site to test whether they are responsible for major cognitive functions like language and memory. This is really important because everybody’s brain is a little different and if a part of the patient’s brain that is responsible for say, forming new memories, overlaps with the epileptogenic site, the neurologists can modify the surgery plan to avoid negatively affecting the patient.

Perhaps the craziest part of this process is that these patients have to be awake while the neurologists assess these other cognitive functions with and without stimulation. The only way to be confident that the patient can still use language after parts of their brain are removed is to have them perform various tasks that they have to be awake to do—like reading or talking— while electrically stimulating these areas.

You might think it would be super painful to have a bunch of metal wires running an electrical current through your brain while you are conscious, but it actually isn't. We can sense pain in lots of the tissue in our head but the brain itself doesn't have any pain receptors, so surprisingly, stimulating most areas isn't painful. I say ‘most’ only because there are just a few areas that receive signals from neurons that are sensitive to pain, so stimulating these select areas could actually induce sensations of pain. It's super interesting and I could go on but we’re quickly getting off topic.

We work closely with lots of neurologists at Northwestern to do experiments with epilepsy patients. It goes without saying that we are extremely aware of the ethical considerations we must take with these patients and everything we do is approved by an internal ethics review board. 

Stimulating the central amygdala (left) but 

not the hippocampus (right) caused patients

to stop breathing.

Members of my lab recently published an experiment that they conducted with epilepsy patients to test this hypothesis—that directly stimulating the central amygdala causes people to stop breathing. It turns out that their hypothesis was correct; every single patient that was tested instantly stopped breathing when the central nucleus of the amygdala was stimulated. We think this is specific to the central amygdala because stimulating neighboring regions had no effect on breathing. Interestingly, breathing only stopped when the patients were instructed to breathe naturally through their nose, not their mouth. There were several fascinating things we noticed. For one, patients stopped their breathing for as long as the stimulus was applied and sometimes longer. However, they were able to resume breathing during stimulation if they were instructed to. Also, (anecdotally) all but one of the patients were totally unaware that they had stopped breathing at all! This is all really exciting and builds on the findings of others [5] [6] but, like most science, raises more questions than it answers. 

The Amygdala and Breathing?

The amygdala has been well-defined as a brain area that is crucial for a lot of emotional processes. If you get rid of someone's amygdalae (the plural of amygdala that nobody uses), they exhibit hypoemotionality, or a lack of emotions. These people can recognize if something is meant to be scary but they don't experience the fear associated with it or exhibit fear-induced behaviors like sweating or elevated heart rate. There have been numerous studies that have shown that people with anxiety disorders, depression, and PTSD all seem to have differences in their amygdalae compared to healthy controls. 

So why the heck is this emotional brain region talking to respiratory areas in the brainstem in the first place? 

We know that when we're emotional, we breathe very differently. For example, laughing and crying are respiratory behaviors that are induced by specific emotional states. Laughing and crying are both so freaking weird and just get weirder the more you think about them. I'm sure that you've laughed so hard that you couldn't breathe, likewise with crying, but hopefully far less often. But that doesn't answer the question. Why is emotion so tied to breathing and how would this be helpful at all? If you were building a brain from scratch, why would you ever give this emotional brain region the authority to override and halt a vital process that keeps you alive? 

I wanted to get my head around this brain region a little better so I asked if I could sit in on my program's neuroanatomy lab. Something I didn't realize is just how tiny the amygdala is! 

In the picture on the left, I'm holding a section of a medical donor's right temporal lobe, which I have sliced coronally. (or at least attempted to)

In picture on the right you can see that the amygdala is only visible in two of these ~1 cm slices. 

Why does this glob of cells, that is roughly the size and shape of an almond, have so much control when we have one of the largest brains in the animal kingdom?

It’s possible that this connection between the amygdala and breathing is some vestigial circuit leftover from evolution or development. A colleague suggested that pausing breathing could help you avoid being detected by a predator, which is very plausable. But maybe, just maybe, the reason that this connection exists might have something to do with smell.


More About The Amygdala and Also Sniffing

"The amygdala is the emotion part of the brain", is a platitude that I've read more times than I can remember. While this is mostly true, I think that the amygdala may be more accurately described as the brain region responsible for determining the hedonic value of stimuli and triggering adaptive responses. In English, that means that when you sense something, your amygdala determines whether it is good for you or bad for you. If you present people with pleasant stimuli, their amygdalae activate. If you show them images of landscapes, or play simple tones, the amygdala doesn't do much. If you show people scary images, their amygdalae go nuts. The amygdala has bidirectional connections with tons of brain areas including the sensory cortices, prefrontal cortex, hippocampus, basal ganglia, hypothalamus, brainstem, and other areas as well. It is like your own personal Central Intelligence Agency that is always gathering intel from all over the brain and monitoring whether things are OK. If your amygdala detects a threat, it sends signals to activate areas of the brain that do things like potentiating reflexes and releasing hormones, which help you react quickly and avoid most kinds of immediate danger [7].

Something that few people know about the amygdala is that it receives extensive, direct input from the olfactory bulb - which is the first area in the brain that processes information about odors. When you smell something awful your amygdala activates like it would if you saw or heard something aversive. However, if you smell something unpleasant you also reflexively stop sniffing [8] [9]. Most things that smell bad are also harmful to you, like rotten food or noxious gasses. Rapidly halting your nasal breathing when you detect potentially dangerous chemicals in your nose prevents you from breathing them into your lungs, where they could be absorbed into your body and harm you. This reflex is key because it demonstrates that olfactory information can rapidly halt ongoing breathing rhythms. This reflex is extremely fast, which has led some researchers to hypothesize the existence of an 'olfactomotor' circuit ('olfacto' meaning olfaction and 'motor' referring to respiratory motor neurons) [10]. The anatomy of the olfactomotor circuit not been investigated yet but I'm bringing this up here because this behavior has amygdala written all over it.

Now I feel like I need to recap what I've told you so far because I've introduced a bunch of pretty tricky ideas in neuroscience that relate to each other in ways that I maybe didn't chain together as well as I could have.

The Story So Far

1. People with severe epilepsy can die from SUDEP.
2. SUDEP seems to be caused largely by fatal cessation of breathing.
3. For a seizure to cause changes in breathing, it must affect the activity of neurons in the respiratory areas of the brainstem.
4. The central amygdala is located in the temporal lobe, where seizures resulting in SUDEP tend to originate, and directly connects to neurons in the brainstem that control breathing.
5. Activating neurons in the central amygdala, both with direct electrical stimulation and seizure activity, causes people with epilepsy to stop breathing.
6. We have a circuit that links information about olfactory hedonic value to respiratory control but its location is unknown.

Here's is the kicker: it seems like the central nucleus of the amygdala is poised to be a key node in this olfactomotor circuit. If this is true, the reason that people stop breathing during a seizure could be because the seizure reaches the central nucleus of the amygdala, activating the olfactomotor circuit, which disrupts nasal breathing. 

So in an ironic and deeply dark way, it could be the case that the circuit that protects us from breathing in harmful chemicals by halting our nasal breathing can be 'hijacked' by a seizure, causing breathing to stop for as long as the seizure persists - potentially leading to SUDEP. 

It's important to point out that these last two statements are theories, not facts. It would be so awesome if we could prevent SUDEP by electrically suppressing activity in the central amygdala. Preventing SUDEP could even be as simple as telling someone who is having a seizure to just breathe through their mouth. However, when you do science, your ideas are wrong far more often than they are right. In fact, it's kind of incredible that we're ever right. Nature is confusing as hell and we're just this weird, emotional primate species trying to figure it all out. The best way to test whether a theory is correct is, un-intuitively, to conduct experiments designed to prove that your theory is wrong. In other words, my research is all about investigating these ideas but every day at work my lab mates and I try to prove ourselves wrong.

These days we're running experiments to see if the central amygdala is not part of the olfactomotor circuit. Specifically, we're trying to identify what chemosensory properties of odors activate this circuit, whether activating the central amygdala actually leads to activation of the brainstem respiratory groups, and how the trigeminal system (something I haven't even mentioned here) might participate in these processes as well. If we successfully fail at proving ourselves wrong, the next steps will be working on treatments for how to use this knowledge to intervene and hopefully prevent SUDEP altogether.


Final Thoughts

I came into grad school almost exclusively interested in brain-computer interfaces, the benefits of which are incredible and obvious. But three years in, I find myself in the hospital, wearing a lab coat, politely asking people undergoing epilepsy surgery to smell a bunch of disgusting odors for me. It's certainly not what I had expected to be doing but I think that I'm a lot happier doing this compared to just about anything else. I think these ideas are fascinating, important, and worth a whole lot of my time to work on, even if I someday have the misfortune of proving all of them wrong. 

...I also really hope that whoever is reading my 64 page grant right now agrees. 



Thanks for reading and I hope that the rest of your day is filled with deep, healthy sniffs of exclusively pleasant odors.